Creating Your First Task

If you’re reading this, you’ve probably logged in to Datax Task Manager for the first time. Task Manager is your access point to Datax Workforce, which consists of researchers like you as well as our extensive network of contributors.

With that in mind, welcome! You’ve just connected yourself with a system of minds that will make your data collection work quicker, cheaper, and more focused. This guide will show you how to create your very first task.

For the purposes of this guide, we’ll be referring to an example task: using our text-to-speech task template to collect audio recordings of text. But don’t worry - the information you find in this guide is widely applicable and will come in handy no matter what task you’re creating.


Before creating your task, you must prepare the resources necessary for our contributors to complete that task.

For our audio collection example, that would be a file containing all the text you need recorded. If you’re using a CSV file, make sure the file has a header row and that you remove any columns not containing text that needs to be recorded - we recommend this because it allows the file to upload faster.

If you have a line-separated text file, just add a new line at the top of the file that reads text, and save a copy as a .csv file.

You’ll upload this file when you reach the Resources page.

Of course, the resources needed for each task will be different. Worry not - if you’re unsure of how to prepare your resource file, you can check out our more detailed guides for specific task types, or contact us for help!

Create a new task

To get started, click New Task on your dashboard, then select the task template you’d like to use, in this case “Text To Speech.”

After this, you can edit some basic information about the task, such as title, cover description and task language.

You're creating a task with many jobs. One line of text is a job, to be distributed to many contributors.

Write a meaningful cover description. Your cover description will attract contributors to complete your task. The more attractive it is, the more responses you’ll be able to choose from.

See Task Cover for more information

Write effective instructions so that contributors will give responses that fit your expectation.

See Task Instructions for more information

Use the resource

Under Data & Resources › Resources, upload your CSV file.

Go back the Design page. Click on the question and then click Add Resource on the right hand side.

A pop-up selection will show the columns of the CSV file, which you can then use to create resource groups, which is how the resource will be presented in each job.

Before publishing the task, you’ll also have the opportunity to set contributor requirements and distribution settings. Both of these sections are quite straightforward.

Distribution settings

Not every response will meet the requirements of your task. Once they start coming in, you’ll have to review them under Responses. Distribution settings and the review process are closely tied. Under Distribution, you are able to control the task’s Response acceptance method. Choosing “Automatic” will accept every response, meaning even poor responses will receive a reward. It is suitable if you are looking for opinions instead of absolute answers. This guide assumes that you choose “Manual,” which allows researchers to review responses one-by-one and only accept and pay for useful ones.

Under Distribution, you should also have been able to set a consensus size, or minimum number of people to complete a job. This should generally be higher for more controversial jobs. Once their consensus size has been met, jobs will be displayed under “Ready for review.” These should be reviewed first.

Publish task

Under Review & Publish, you should find instructions to get the task published. Most settings are not editable once the task is published, so make sure you have made all necessary changes.

Review responses

We recommend reviewing responses this way: Reject problematics ones, then accept the rest at once.

Reject only responses that you cannot use to speed up the completion of your task; responses are impossible to export before being accepted. When you reject a response, be sure to select the right reason, as this feedback helps the Datax team provide better guidance and training to contributors.

Good luck and don’t hesitate to contact us with any questions!